A computer-implementable method for managing a cognitive graph comprising: receiving data from a plurality of data sources; processing the data from the plurality of data sources, the processing the data from the plurality of data sources identifying a plurality of knowledge elements; and, storing the knowledge elements within the cognitive graph as a collection of knowledge elements, the storing universally representing knowledge obtained from the data, each knowledge element being structurally defined within the cognitive graph.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A computer-implementable method for managing a cognitive graph of a universal knowledge repository comprising: receiving, by an information processing system that includes a processor, data from a data source; processing, by a cognitive inference and learning system executing on the information processing system, the data from the data source, the processing of the data from the data source identifying a plurality of knowledge elements, the plurality of knowledge elements representing quantity in a universal cognitive graph of the universal knowledge repository, the cognitive inference and learning system comprising a cognitive platform, the cognitive platform comprising the universal cognitive graph: storing, by the information processing system, the knowledge elements within the universal cognitive graph as a collection of the knowledge elements, the storing universally representing knowledge obtained from the data, each knowledge element of the knowledge elements being structurally defined within the universal cognitive graph, the collection of knowledge elements being accurately and precisely modeled via a ladder entailment pattern and a diamond entailment pattern, the ladder entailment pattern comprising a pattern formed in the universal cognitive graph where rungs of the ladder entailment pattern are formed by a first set of nodes having a categorical relationship with a corresponding second set of nodes and rails of the ladder entailment pattern are formed by attributive relationships respectively associated with the first set and second set of nodes, wherein the nodes are the entities, and wherein structurally defining each knowledge element being implemented with an ontology, the ontology universally representing knowledge and comprising a representation of the entities along with properties and relations of the entities according to a system of categories, the ontology storing a knowledge element within the universal cognitive graph based upon a set of the categories of the knowledge element and a set of attributes of the knowledge element; receiving, by the information processing system from a user via a network, a user query that includes text of the data to generate a graph query; performing, by the information processing system based on mapper rule set, mapping operations on the user query; parsing, by the information processing system based on using a parsing operation or a parse rule set, the user query to identify ambiguities of the text after the performing mapping operations, the parsing operation comprising a lossless parsing operation, the lossless parsing operation identifying a minimum basic known true structure for all parse variations, the using of the parse rule set resulting in generation of a set of parse trees, each parse tree of the parse trees being represented as the text being parsed; ranking, by the information processing system based on the parse ranking rule set, the parse trees which result in determination of top-ranked parse trees to resolve ambiguities in the text of the query, the parse trees being transformed to a tree representing an interpretation of ambiguous portions of the data, the interpretation of the ambiguous portions of the data being changed based upon user Feedback, the top-ranked parse trees are stored in the universal knowledge repository; identifying, by the information processing system, knowledge elements of the parse trees that represent ambiguous portions of the text, identified knowledge elements of the parse trees being stored within the universal cognitive graph in a configuration representing a relationship of concepts; querying in response to the user query, by the information processing system, the universal knowledge repository and using the identified knowledge elements of the parse trees and the identified plurality of knowledge elements of the universal cognitive graph to produce a plurality of cognitive insights as recommendations that satisfy the graph query; ranking, by the information processing system based information in insight streams, the plurality of cognitive insights that are provided in a cognitive insight summary, the information being related to a location of the user, the user feedback and a device used by the user; publishing, by the information processing system, the ranked plurality of cognitive insights in the cognitive insight summary to the user of cognitive insight data, the ranked plurality of cognitive insights being related to a cognitive user profile.
This invention relates to a system for managing a universal knowledge repository using a cognitive graph to process and query structured and unstructured data. The system receives data from various sources and processes it using a cognitive inference and learning system to identify knowledge elements, which are then stored in a universal cognitive graph. The graph models knowledge using an ontology that defines entities, their properties, and relationships, structured through ladder and diamond entailment patterns. The ladder pattern connects nodes (entities) via categorical and attributive relationships, while the diamond pattern further refines these connections. The system also handles user queries by mapping and parsing the input text to resolve ambiguities, generating parse trees that are ranked to determine the most accurate interpretation. Ambiguous portions of the query are identified and stored in the cognitive graph, where they are linked to existing knowledge elements. The system then queries the repository to produce cognitive insights, which are ranked based on user location, feedback, and device context, and presented in a summary tailored to the user's cognitive profile. This approach enables dynamic, context-aware knowledge retrieval and recommendation.
2. The method of claim 1 , wherein: structurally defining a knowledge element comprises storing a knowledge element within the cognitive graph based upon a set of categories of the knowledge element and a set of attributes of the knowledge element.
This invention relates to knowledge representation and management within a cognitive graph structure. The problem addressed is the efficient organization and retrieval of knowledge elements in a structured, attribute-based system. The method involves defining and storing knowledge elements within a cognitive graph by categorizing them based on their inherent characteristics and associated attributes. Each knowledge element is assigned a set of categories that describe its broader classification, such as subject matter or domain, and a set of attributes that provide specific details about the element, such as properties, relationships, or metadata. These categories and attributes are used to determine the element's position and connections within the cognitive graph, enabling structured storage and retrieval. The cognitive graph serves as a dynamic, interconnected network where knowledge elements are linked based on their categories and attributes, facilitating efficient querying, analysis, and inference. This approach enhances knowledge management by allowing precise organization and retrieval of information based on both high-level classifications and fine-grained attributes. The method ensures that knowledge elements are stored in a way that preserves their contextual relationships, improving the accuracy and relevance of retrieved information.
3. The method of claim 1 , wherein: a relationship between a first knowledge element and a second knowledge element comprises only one of a categorical relationship and an attributive relationship, the categorical relationship representing inheritance of an entity and the attributive relationship representing attribution of the entity.
This invention relates to knowledge representation and management systems, specifically addressing the challenge of defining and organizing relationships between knowledge elements in a structured and unambiguous manner. The method involves establishing a relationship between a first knowledge element and a second knowledge element, where the relationship is strictly limited to either a categorical relationship or an attributive relationship. A categorical relationship signifies inheritance, indicating that one entity is a subclass or subtype of another, while an attributive relationship denotes attribution, meaning one entity possesses or is associated with certain properties or characteristics of another. By restricting relationships to these two distinct types, the system ensures clarity and consistency in knowledge modeling, preventing ambiguity and logical inconsistencies that may arise from overlapping or undefined relationship types. This approach enhances the precision of knowledge representation, making it easier to query, analyze, and derive insights from structured knowledge bases. The method is particularly useful in domains such as artificial intelligence, semantic networks, and expert systems where accurate relationship definitions are critical for reasoning and decision-making processes.
4. The method of claim 1 , wherein the diamond entailment pattern comprises a pattern formed in the universal cognitive graph when two parent nodes of a first node inherit from a common second node.
The invention relates to cognitive graph analysis, specifically the identification of diamond entailment patterns within a universal cognitive graph. The universal cognitive graph is a structured representation of knowledge where nodes represent concepts and edges represent relationships between them. A diamond entailment pattern occurs when two parent nodes of a first node both inherit from a common second node, forming a diamond-shaped structure in the graph. This pattern is significant because it indicates a logical or semantic relationship where the first node is influenced by two distinct but related parent concepts that share a common ancestor. The method involves detecting such patterns to infer hierarchical or entailment relationships within the graph, which can be used for knowledge extraction, reasoning, or decision-making applications. The detection process may include traversing the graph to identify nodes and their relationships, analyzing inheritance paths, and confirming the diamond-shaped structure. This approach helps uncover implicit knowledge structures that may not be explicitly defined in the graph, improving the accuracy and completeness of cognitive reasoning systems. The invention is particularly useful in artificial intelligence, semantic web technologies, and knowledge management systems where understanding complex relationships between concepts is critical.
5. The method of claim 1 , wherein: inheritance relationships between a base node and a parent node define a particular concept within the universal cognitive graph.
The invention relates to a system for organizing and representing knowledge within a universal cognitive graph, addressing the challenge of efficiently modeling complex relationships between concepts. The method involves defining inheritance relationships between nodes in the graph, where a base node represents a fundamental concept and a parent node represents a derived or specialized concept. These relationships establish hierarchical structures that capture the semantic connections between concepts, enabling reasoning and inference within the graph. The inheritance relationships may include properties such as subsumption, where the parent node inherits attributes or behaviors from the base node, or specialization, where the parent node refines or extends the base node's definition. The graph may also incorporate additional nodes and relationships to further refine the representation of concepts, such as associative links between unrelated concepts or temporal dependencies. The system allows for dynamic updates to the graph, enabling the addition, modification, or removal of nodes and relationships as new knowledge is acquired or existing knowledge is refined. This approach facilitates scalable and adaptable knowledge representation, supporting applications in artificial intelligence, semantic search, and cognitive computing.
6. The method of claim 1 , wherein: a plurality of knowledge elements representing different quantities are stored within the universal cognitive graph.
A system and method for organizing and retrieving knowledge elements within a universal cognitive graph. The technology addresses the challenge of efficiently storing, linking, and retrieving diverse knowledge elements representing different quantities, such as numerical values, categorical data, or abstract concepts, in a structured yet flexible manner. The universal cognitive graph serves as a unified framework where knowledge elements are stored as nodes, with relationships between them represented as edges. This structure enables semantic reasoning, contextual understanding, and dynamic querying across heterogeneous data types. The method ensures that knowledge elements are uniquely identifiable and interoperable, allowing for scalable knowledge integration and retrieval. By leveraging graph-based representations, the system supports complex queries, inference, and knowledge discovery, improving decision-making and problem-solving in domains requiring multi-faceted data analysis. The approach enhances traditional knowledge management systems by providing a more adaptive and interconnected representation of information.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
February 11, 2020
April 5, 2022
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.